Skip to content

Latest commit

 

History

History
38 lines (28 loc) · 2.16 KB

glitch-soc.md

File metadata and controls

38 lines (28 loc) · 2.16 KB

Enabling Glitch-Soc

Lysand supports the use of the Glitch-Soc fork of Mastodon's frontend. Here's how to do it:

  1. Download the latest Glitch FE package from the releases page (it should be named glitch.tar.gz and be inside the assets of a normal Lysand release).

  2. Extract the contents of the package into a folder, which you can name glitch for simplicity. (if using Docker, now is the time to mount that folder into the container, for example with -v /path/to/glitch:/app/dist/glitch)

  3. Change the config to enable Glitch-FE:

    [frontend]
# Enable custom frontends (warning: not enabling this or Glitch will make Lysand only accessible via the Mastodon API)
enabled = false
# The URL to reach the frontend at (should be on a local network)
url = "http://localhost:3000"

[frontend.glitch]
# Enable the Glitch frontend integration
enabled = true
# Glitch assets folder
assets = "glitch"
# Server the assets were ripped from (and any eventual CDNs)
server = ["https://tech.lgbt"]

    (you can disable the normal frontend option as it will not be used anymore)

    The server option can be left as-is, unless you have downloaded your own index.html file from a different Glitch instance.

  4. Start Lysand and navigate to / to see the Glitch frontend in action.

How is this package created?

Glitch-FE is a React single-page app, which dynamically creates an index.html file on every request and builds all the other assets at build time.

The package static files were taken from a build of Glitch-Soc, while the index.html file was taken from the tech.lgbt instance with cURL.

Then, the paths in the index.html file were replaced with the correct paths for the static files (as they have different hashes in their names from the ones in the Glitch-Soc build).

At runtime, Lysand dynamically edits the index.html file to replace the content with correct data, as well as disabling integrity checks on the script/link tags.

In the future, I'll find a way to make this less hacky and more user-friendly, but for now, this is the best I can do.